The disease course of multiple sclerosis (MS) varies greatly among individuals, making prognostication difficult yet crucial. Although recent decades have brought advancements in the diagnosis and treatment of MS, integration of reliable prognostic markers into clinical practice remains an ongoing effort.
Serial magnetic resonance imaging (MRI) scans and clinical assessments remain the standard for evaluating disease activity and, thus, the main considerations guiding therapeutic decision-making. However, these methods have limitations: they are resource-intensive and impractical for frequent use. This highlights the need for easy-to-obtain, frequent, and objective biomarkers that can accurately track disease activity, helping to identify people who require higher-efficacy treatment or more intensive monitoring.
In recent years, multiple prognostic biomarkers have been studied. Of these, neurofilament light chain (NfL) protein is one of the most promising. In this narrative review, we discuss the current evidence behind the use of NfL as a prognostic biomarker in MS and the current barriers and remaining unresolved questions that must be addressed before its broader adoption in clinical settings.
Neurofilament Light Chain
Neurofilaments are neuron-specific cytoskeletal components that play a large role in maintaining axonal function, transport, and homeostasis when fully intact. Neurofilaments consist of three subunits—a light, medium, and heavy chain—which can be detected individually after axonal damage. Of these subunits, NfL protein has been established as the most easily detectable and reliably related to axonal damage in MS.1
Early studies of NfL were hindered by the fact that it could only be measured through cerebrospinal fluid (CSF), as it exists in much lower concentrations in the blood.2 Initial assays used for the detection of NfL in the serum (sNfL) and plasma were developed in the early 2000s; studies using peripheral blood measurement began to increase by 2016.2 Since then, there has been substantial growth in the number of studies dedicated to understanding the role of NfL as a marker of axonal damage not only in MS but also in many other neurologic diseases.
From a practical standpoint, NfL can be measured in both plasma and serum. Studies comparing CSF, plasma, and serum NfL have shown that all three methods correlate well.2,3 However, most recent studies have performed serum measurement and analysis using the Single Molecule Array (SiMOA)-based assay, largely because this was one of the first assays that was able to detect NfL in extremely low concentrations, down to nanogram and picogram levels.2,3 Detection methods in both serum and plasma continue to evolve, with new proprietary assays emerging using proteomics and high-resolution mass spectroscopy to improve specificity, interassay variability, and cost. Because both serum and plasma assays are in use, all mentions of NfL in this review are in relation to plasma and serum measurement of NfL, unless stated otherwise.
Factors Influencing the Interpretation of NfL
NfL is released in any state involving axonal damage, therefore making it a nonspecific marker of neurodegeneration. Table 1 highlights some of the main confounding factors that must be considered when interpreting measurements of NfL.
Timing of NfL Measurement. The timing of NfL measurement in relation to clinical or sub-clinical disease activity is known to be important and to greatly affect NfL levels. Thus, interpreting a single NfL measurement can come with uncertainty. In clinical practice, the value of NfL might lie in monitoring trends in individuals, which can account for varying factors on a personal level. Such changes can be useful in identifying key milestones in MS course, such as predicting a relapse, spotting early progression, or confirming how well a treatment is working. In the following, we discuss the current research as it pertains to the changes in NfL throughout the natural course of MS.
NfL as a Prognostic Biomarker in the Preclinical Phases of MS
Major therapeutic advancements have led to a push to identify and risk-stratify asymptomatic individuals in the preclinical stages of MS. Current studies are attempting to identify and characterize asymptomatic and early phases of MS and investigate biomarkers for conversion from a clinically isolated syndrome (CIS) or radiographically isolated syndrome (RIS) to clinically definite MS (CDMS).
Current data suggest that NfL is elevated in apparently healthy individuals who will go on to develop CDMS. In a landmark case–control study using sNfL measurements in US military personnel from 2000 through 2011, Bjornevik et al7 demonstrated that sNfL was elevated in the 60 individuals who would go on to develop CDMS. There was a temporal component to this NfL elevation: NfL levels were increased an average of 6 years before the diagnosis of CDMS, and increased serially as the individuals approached their first clinical attack.7 These data indicate that neuroaxonal injury begins early and can precede any clinically detectable symptoms.
NfL levels are higher in individuals with CIS or RIS compared with age-matched controls, and correlate with markers of disease activity, including MRI changes (i.e., T2 lesions, gadolinium-enhancing lesions, diffusely abnormal white matter) and Expanded Disability Status Scale (EDSS) scores.2,8,9 Some studies, however, have not been able to replicate this relationship.10 These equivocal findings may be related to smaller sample sizes as well as the temporal dynamics between NfL and time to a clinical attack.
NfL as a Prognostic Biomarker Throughout the Natural Course of MS
Early studies have had conflicting results, especially when analyzing the relationship between NfL and markers of clinical progression or disability accumulation (e.g., EDSS scores, cognitive decline, depression, fatigue).9-11 Whereas methodologic considerations and sample size may have contributed to conflicting results, the timing of NfL measurement likely plays a large role. Later studies have explored the relationship between NfL levels with different time points in the natural course of MS and have yielded some promising results, which are summarized below.
NfL as a Potential Predictor of Impending Relapse
Recent studies have independently demonstrated that elevated baseline levels of NfL are associated with an increased risk of impending relapse and the development of new gadolinium-enhancing lesions.12-18 NfL was shown to be increased by 32.3% before and during clinical and radiographic relapse, and then normalized in remission.12
Although studies have shown consistently that NfL levels can be used to predict relapse, the estimated time until relapse will occur has differed, likely because of different study designs. At least three studies showed an increased risk of clinical or radiographic relapse within three months12-14; however, other studies showed an increased risk of impending relapse within one and two years.15,16
These studies also differed in terms of the cut-offs used to calculate the risk of impending relapse. Two studies used absolute sNfL values, two studies used differences between serial measurements, and two studies used z scores, a measure that quantifies how far a data point is from the mean, in terms of standard deviations, adjusted for age and body mass index (BMI).12-17 When absolute sNfL measurements were used, these studies differed in established cut-off values. Despite using the same SiMOA-based assays, Brune et al16 used a cutoff value of 8 pg/mL, whereas Cutter et al14 used 16 pg/mL plus one gadolinium-enhancing lesion as a predictor of impending relapse.
Altogether, the studies suggest an exciting opportunity for identifying individuals in need of more aggressive monitoring and potential treatment escalation. However, more real-world data needs to be generated to assess the timing of impending relapse relative to NfL increases to validate precise NfL monitoring intervals and cut-offs.
NfL as a Potential Predictor of the Onset of Progression Rather Than Disability Accumulation
Previous studies assessing the relationship between NfL and measures of disability accumulation have been equivocal. A recent systematic review9 demonstrated that NfL and EDSS scores were correlated in four studies, but not in six others. Studies comparing NfL and other markers of clinical progression, such as the 25-foot walk test or neuropsychological assessments have had equally conflicting results.19
Given the large number of studies that have been performed, this is unlikely to be entirely related to methodologic differences, but likely reflects true variability. The limited sensitivity of existing clinical scales for measuring disease progression could account for some of the negative findings as by the time results on these scales become abnormal, a progressive neurodegenerative pathology may have been ongoing for years.
More recent studies suggest that NfL may be a good prognostic biomarker in identifying the earliest biochemical signs of the switch from a more peripherally mediated inflammatory milieu to that of progressive CNS-compartmentalized neurodegeneration. A 2021 study by Uphaus et al20 was one of the first to suggest this, showing that elevated levels of sNfL at baseline and persistently elevated levels at follow-up were associated with relapse-free progression in EDSS scores and eventual conversion to secondary progressive MS by the six-year mark, respectively. A more recent study by Abdelhak et al15 showed similar findings, this time demonstrating that NfL increases preceded clinical progression by one to two years.
These studies shed light on NfL as a potential predictive biomarker, in which abnormalities precede clinical measurement of disability progression by years. This suggests that the axonal damage measured by NfL occurs early, and by the time conventional measures of disability progression and accumulation become abnormal, much of the axonal damage has already been done. Whereas the potential of NfL as an early biomarker of progression is interesting, more studies are needed to further characterize and quantify this complex relationship.
NfL as a Potential Indicator of Treatment Response
NfL is also emerging as a potential marker for assessment of treatment efficacy. The existing data have been concordant, with a large body of evidence dating to the CSF studies of the early 2000s. These studies showed that NfL levels decreased reliably after initiation of disease-modifying treatment (DMT) and remained increased without treatment. These post treatment decreases in NfL also correlated with other surrogate markers of response, such as fewer gadolinium-enhancing lesions, relapses, and other MS symptoms.14,18,21
More recent studies have shown that NfL also has a graded response to low-, moderate-, and high-efficacy DMTs. Studies consistently demonstrate substantial decreases in NfL levels after treatment with high-efficacy monoclonal antibodies (natalizumab, alemtuzumab, ocrelizumab, rituximab), less-pronounced decreases with moderate-efficacy therapies (teriflunomide, dimethyl fumarate, fingolimod, siponimod), and marginal decreases with low-efficacy therapies (interferons, glatiramer acetate).21,22 Complementary therapies, such as vitamin D supplementation, have also been studied and have shown no effect on either serum or NfL or plasma NfL levels.23
Conclusions and Future Directions
This narrative review explores the potential of blood NfL as a prognostic biomarker that has multiple applications at various time points over the natural course of MS. The most promising current applications lie in its ability to predict relapses or subclinical disease activity and to monitor treatment response. Although certain studies suggest that NfL may help predict the transition to a more progressive form of MS, further research is necessary to better understand this relationship.
Advancements are being made in the development of new measurement technologies.24 Such new technologies are expected not only to lower the cost of the test but also to enhance its accessibility, facilitating more frequent testing and thereby enriching patient care and monitoring. It is imperative that additional real-world data be gathered to validate the findings from research studies on NfL outlined in this review.
Despite a need for further studies, the future of NfL as a prognostic marker for surveillance of disease activity and treatment response is bright, and presents an exciting, objective, and easily obtainable method of providing personalized management for people with MS.
